Copy number variation (CNV) is a common genetic structural variation that contributes greatly to human genetic diversity. CNVs are defined as deletions and duplications of genomic DNA segments that are at least one kilobase long . Thousands of CNV loci have been identified throughout the human genome . Several mechanisms have been proposed for the origination of CNV, including nonallelic homologous recombination, nonhomologous end-joining, and fork stalling and template switching . The genomic rearrangements responsible for CNVs can lead to pathogenic phenotypes by modulating gene dosage, interrupting a gene, creating a fusion gene, exerting position effects, or unmasking a deleterious recessive mutation . In fact, CNVs in specific genes have been linked to a variety of diseases, including Parkinson's disease, HIV infection, psoriasis, systemic lupus erythematosus, and a variety of neuropsychiatric diseases [42–46]. Nevertheless, the role of CNV in AMD remains incompletely understood.
It has been reported that deletions in CFH-related proteins 1 and 3 (CFHR1 and CFHR3) are protective against AMD. CFH and CFHR1–5, regulators of complement activation, are located in a common locus in linkage disequilibrium on the long arm of chromosome 1. Hageman et al. identified a large deletion spanning CFHR1 and CFHR3 and demonstrated that homozygosity for this deletion conferred a protective effect, which was particularly evident for advanced AMD . Hughes et al. identified this same deletion in two independent cohorts and demonstrated that its protective effects could not be attributed to the CFH Y402H genotype . It was later demonstrated that its effects are also independent of the risk-modulating SNP in the CFH promoter region . A previously characterized CFH haplotype, which is protective against AMD, occurs frequently in homozygotes for the CFHR1 and CFHR3 deletion. This copy number polymorphism may indeed be responsible for at least part of the protection conferred by the haplotype . Schmid-Kubista et al. characterized CNV in CFHR1 and CFHR3 in 252 AMD patients and 249 controls using multiplex ligation-dependent probe amplification, which allows for quantitative determination of gene copy number . This study reported novel CNVs in this locus, including homozygous deletions of only CFHR1 or only CFHR3, heterozygous deletion of CFHR3 only, and duplication of CFHR1 only. Homozygous deletion of both genes decreased the odds of AMD substantially, and the protective effects of these CNVs were statistically significant for early AMD and neovascular AMD. The effects of heterozygous deletion did not achieve statistical significance, but a trend suggested a gene dosage effect. It was recently estimated that having fewer than two copies of CFHR1 and CFHR3 reduced the odds of having AMD by 43% . CFHR1 and CFHR3 are regulators of the complement system that inhibit C5 convertase and C3 convertase, respectively. They are also competitive inhibitors of CFH and interrupt the binding of CFH to C3 . A recent report suggests that the protection conferred by deletion of CFHR1 and CFHR3 may not be independent of rs1329428 and rs203687, two SNPs downstream of CFH Y402H in a novel 32-kb region associated with AMD risk . This study also showed that a different deletion CNV in a 122-kb region encompassing CFHR1 and CFHR4 was associated with protection against AMD, independently of other CFH SNPs. The demonstrated role of CNVs and SNPs in modulating the activity of genes in the regulator of complement activation locus further highlights the mechanistic importance of the complement system in AMD pathogenesis.
CNVs in the glutathione S transferase (GST) genes are common in human populations, and their role in a variety of diseases has been investigated. GSTs are phase II enzymes that defend against oxidative stress and detoxify a variety of electrophilic compounds by covalent conjugation with glutathione . Homozygous deletion of GSTM1 and GSTT1 occurs in approximately 50% and 20% of Caucasian individuals, respectively . A decrease in GSTT1 copy number has been linked with cortical cataract , and deletions of GSTM1 and GSTT1 are associated with asthma . Preliminary investigations have been performed to evaluate the potential role of GST CNV in AMD. In a study by Oz et al., copy number genotyping was performed for GSTM1, GSTT1, and GSTP1 in 35 patients with neovascular AMD and 159 controls, and no statistically significant associations were observed between GSTM1, GSTT1, or GSTP1 individually and AMD . However, homozygous deletions for GSTM1/GSTT1 and GSTM1/GSTP1 in combination were both found to significantly increase AMD risk. Guven et al. evaluated GSTM1 and GSTT1 CNV in 120 AMD patients and 198 controls in a Turkish population . Homozygous deletion of GSTM1 was found to be associated with AMD, and the association persisted for dry AMD after stratification by AMD subtype. Statistically significant associations were not observed between GSTM1 or GSTT1 genotype and neovascular AMD. A study by Kimura et al. in a Japanese population also reported no significant association between CNV in GSTM1 or GSTT1 and neovascular AMD .
The potential role for CNV in AMD has been evaluated for a number of other genes. A recent genome-wide scan for CNVs was performed using the Affymetrix GeneChip SNP Microarray in 400 AMD patients and 500 elderly controls. Nephrocystin 1 (NPHP1) and EGF containing fibulin-like extracellular matrix protein (EFEMP1) were identified as genes of interest . Mutations in the NPHP1 gene have been reported in Senior-Loken syndrome, which occurs with Leber congenital amaurosis, in Joubert syndrome, which may cause retinal dystrophy, and also in juvenile nephronophthisis in isolation. A deletion on chromosome 2q13 containing NPHP1, which functions in signal transduction and cytoskeleton organization, was observed in AMD patients, but not in controls. Duplication of a region upstream of EFEMP1, a gene that is mutated in Doyne honeycomb retinal dystrophy, was also observed in AMD patients, but not in controls. The reported CNVs were quite rare, and their mechanistic implications in the context of AMD remain entirely unknown and require further investigation.
We also recently evaluated the potential role of CNV in neovascular AMD for several candidate genes, including CCR3, CFH, CX3CR1, ERCC6, HTRA1, VEGF, GSTM1, and GSTT1. Quantitative copy number genotyping was performed for each gene in 131 neovascular AMD patients and 103 elderly controls. Previously unreported CNVs were discovered in CCR3, CX3CR1, and ERCC6, but after adjustment for age, no statistically significant associations were found between CNVs in any of the evaluated genes and AMD. A trend in the unadjusted data suggested that an increase in CX3CR1 copy number might be protective against AMD. This finding complements previously reported studies characterizing the association between two loss-of-function SNPs in CX3CR1, V249I and T280M , and elevated AMD risk. Additionally, histological studies have demonstrated that macular CX3CR1 protein levels are decreased in patients with AMD . Further evaluation of the role of CNV in AMD is warranted to better understand the contribution of this type of genetic variation to AMD risk and pathogenesis.